Customer Segmentation Analysis
Completed:
Problem Statement
Businesses need to understand their customer base to develop targeted marketing strategies. Manual segmentation is inefficient and often misses patterns in high-dimensional data. This project applied unsupervised learning techniques to automatically identify distinct customer groups based on demographic and behavioral variables.
Approach
- Exploratory Data Analysis: Conducted comprehensive EDA on a 2,217-record marketing dataset
- Statistical Testing: Performed χ²-test and Correspondence Analysis to examine relationships across demographic and behavioral variables
- Dimensionality Reduction: Applied Principal Component Analysis (PCA) to reduce feature space while retaining 95%+ variance
- Clustering: Used K-Means clustering algorithm to identify optimal customer segments
- Interpretation: Analyzed cluster characteristics to extract business insights on income and purchasing behavior
Tech Stack
Python pandas NumPy SciPy Scikit-learn Matplotlib
Key Insights
- Identified two distinct customer segments with differentiated income and purchasing behavior patterns
- Income and purchasing behavior emerged as primary drivers of customer differentiation
- Results provide actionable segments for targeted marketing campaigns and resource allocation
Presentation
Date: January 22, 2026
Dataset Size: 2,217 records
Status: Completed